AgOpenGPS

JohnBoy

Well-Known Member
Theres been a number of mentions of this project recently and I thought it might be worth starting a thread on it.

I've been reading a lot about it and am going to experiment a bit with it.

For those who haven't heard of it AgOpenGPS is an open source project to create an agricultural guidance system that is non proprietary and cheap. With the commodisation of GPS technology, combined with the availability of electronic controllers like arduino it is possible to build many things today for small money that would have cost tens of thousands a few years ago. (Open source us a software term used to describe projects where the source code is openly available and can be contributed to by multiple interested developers)

The AgOpenGPS software runs on a windows tablet or laptop. At a basic level you connect a usb GPS receiver and away you go, it does all your AB lines and guidance stuff and has a light bar drawn on the screen for steering.

That basic setup wont be much good for anything other than spreading fertiliser on flat grassland. And only then if you've spent probably 200 on a very good receiver.

But AgOpenGPS is very very extendable. It can accept an RTK correction over the internet which should make even a cheaper receiver more accurate.

There is a sub project to the software around a circuit board with an arduino controller that can talk to electronic compasses and gyroscopes which will improve accuracy on hills.

The same circuit board can also run an LED light bar as tablet screens are not easy to look at in the sun.

The circuit board is really intended to drive auto steer motors and their are people running it with both wheel motors turning the steering wheel but also in hydrostatic systems that are capable of being steered electronically too.

Another sub project involves another circuit board that joins two GPS receivers together which will allow for extremely accurate heading/roll correction by comparing the relative positions of two nearby receivers.

Its fascinating stuff and is being used to control all sorts of machines.

I thought a thread here might be useful on it. I'll post what I learn as I go, if anyone else does the same it could become a useful resource.


A big issue with this and many other open source projects is the high barrier to entry. The people involved are very advanced users and its all voluntary so the basic how to type guides can be lacking or written at a level that still assumes a lot of knowledge. I'm hoping to gain a bit of knowledge and contribute a beginner's guide to building a grassland grade system as opposed to the full on drilling without tramlines using extreme accuracy level that many are working with this tool for.
 
One of the key requirements for accurate guidance is a correction signal, GPS is inherently inaccurate. Theres only so much you can do with a couple of radio waves coming from space.

Truly accurate guidance requires a correction signal, broadcast from a known fixed position, generally within a 20km radius from where you're working. There are a number of companies that offer RTK correction services for a thousand or so per annum, but again this is something that can be done by combining GPS tech with mobile data networks.

With a receiver and a computer in a fixed location one can upload correction data to the internet where it can subsequently be downloaded again by the computer running AgOpenGPS over a mobile hotspot off a phone.

This is the key to this tech being affordable. With a few hundred in gps receivers you can have sub 10cm repeatable accuracy.


This is something that I've yet to find anything about in Ireland as to whether any public body is offering a free service like this. If not I think it would be a great thing for the Co-Ops to do. They have sites all over the country with good internet connections and it would cost very little for them to provide a service like this.

Setting up such a base station is one of my first priorities.
 
The second thing I'm working on is to get a light bar going. It looks very simple and I've bought and arduino nano to play with. As I say the overall project is very auto steer focused but it is capable of running an led light bar which should cost very little to build.

Once I have a correction server and a light bar figured out then I can play with some receivers
 
If I read things right Agopen should be able to run using nmea,this would then mean my steer ready tractor can use can bus signals off agopen to steer,all very interesting stuff but takes ages to read through.
 
If I read things right Agopen should be able to run using nmea,this would then mean my steer ready tractor can use can bus signals off agopen to steer,all very interesting stuff but takes ages to read through.
You can set up your own base station for gps correction Signals at less than 1k and it works with Agopengps
https://www.ardusimple.com/.


really would like to experiment with agopengps but you need a good quality tablet to be able to see it in a tractor. That’s where the commercial gps screen has the advantage. I’ve tried to get my ez guide 500 to accept an external gps input (actually tsip) but had to give up for now anyway.
 
If I read things right Agopen should be able to run using nmea,this would then mean my steer ready tractor can use can bus signals off agopen to steer,all very interesting stuff but takes ages to read through.

NMEA is the language of GPS receivers. It's just your location as opposed to guidance.

AgOpenGPS is taking that in to make decisions on guidance.


I'm not sure how it works guiding a steer ready tractor, I don't think AgOpenGPS is capable of doing it via csnbus, I think it's that it can be used to hijack the steering valve directly. But I'd say that's very advanced level usage of it.
 
You can set up your own base station for gps correction Signals at less than 1k and it works with Agopengps
https://www.ardusimple.com/.

You can use a pair of ardusimple units and their own wireless setup to create a local private RTK with a theoretical 20km range but from reading that range is only in very ideal conditions.

If you're working in mobile coverage the ntrip approach seems to be better where you upload your correction data from your fixed base to www.rtk2go.com and then via mobile you download the correction data to AgOpenGPS.

This has the added advantage of being non proprietary and available to others (open source)

One may even be lucky enough to have someone local doing this already. http://monitor.use-snip.com/map

really would like to experiment with agopengps but you need a good quality tablet to be able to see it in a tractor. That’s where the commercial gps screen has the advantage. I’ve tried to get my ez guide 500 to accept an external gps input (actually tsip) but had to give up for now anyway.

Especially for driving off it, the screen is a major weakness of any pc based option.

That's why I'm like at making the LED light bar as well. Also going to experiment with and anti glare filter and see if it helps.
 
NMEA is the language of GPS receivers. It's just your location as opposed to guidance.

AgOpenGPS is taking that in to make decisions on guidance.


I'm not sure how it works guiding a steer ready tractor, I don't think AgOpenGPS is capable of doing it via csnbus, I think it's that it can be used to hijack the steering valve directly. But I'd say that's very advanced level usage of it.
Looks like he’s doing it here.
 
You can set up your own base station for gps correction Signals at less than 1k and it works with Agopengps
https://www.ardusimple.com/.


really would like to experiment with agopengps but you need a good quality tablet to be able to see it in a tractor. That’s where the commercial gps screen has the advantage. I’ve tried to get my ez guide 500 to accept an external gps input (actually tsip) but had to give up for now anyway.
If I could get a tablet to link into my existing screen using vga it would be a tidier set up,I’m fairly sure it can be done.
 
Looks like he’s doing it here.


Steering via canbus would be a million miles above my level of interest so I don't know, but from reading the descrtiption/comments on that video I think he's just getting the steer on/off button via canbus, not doing the steering that way.
 
Ok, the hardware I'm using for the display and lightbar

Tablet:
I have a Microsoft Surface Pro 4, currently belongs to work, but any windows tablet should do. they're available on ebay second hand from around £100. you don't want anything running an atom processor or "Windows RT" if you're more serious about this then a panasonic tablet would be the way to go, they start from around £200. but for testing purposes any windows laptop will do, just hard to use a trackpad in a tractor is all.

Controller:
I'm using an Arduino Nano clone, this is the controller that the more advanced autosteer uses so it makes sense to use the same thing. this is the one I got:
https://www.ebay.co.uk/itm/Arduino-Nano-Micro-USB-V3-0-Compatible-Optional-Headers-Atmega328-UK-Stock/153462656107?ssPageName=STRK:MEBIDX:IT&var=453416113303&_trksid=p2057872.m2749.l2649
(I ordered it with header pins attached)

Light Strip:
These are referred to as NeoPixels or WS2812 LED strips. I wouldnt reccomend the one I bought as it requires 12v power, they come in 5v versions too which would be easier to control work with as the arduino could power it directly.
Generally sold by the metre and can be cut to length

I've bought a second strip, the first one could only be adressed in groups of three LEDs, so to have a 13 led bar would actually be a 39 led bar and be 2 foot wide!

Have bought this instead. 60 per metre, 5v, IP65 version. coming from china which will take a while but I can continue working with the strip I have, it'll be functionally the same, just bigger.

https://www.ebay.co.uk/itm/5V-WS281...var=502749944231&_trksid=p2057872.m2749.l2649

Breadboard:
Not strictly nessecary but makes prototyping much easier. it's a board you can plug things into to make links to circuits.
I already have one (I've been a nerd a good 30 years now) but something like this would be fine: https://www.ebay.co.uk/itm/Solderle...021414?hash=item3d22b0dd66:g:VtUAAOSwaB5XrgU-

Other stuff:
some jumper cables, resistors, switches and LEDs could be handy for figuring some stuff out.
This kit: https://www.ebay.co.uk/itm/Emakefun...198576?hash=item262982d9f0:g:lUsAAOSwx5VdpH0z looks great. it includes an arduino nano, breadboard and lots of bits and pieces for £18



That's as far as I've gone hardware wise. there's a bit of learning to be able to get that much up and running.

Start with this tutorial to figure out the basics of the code editor and getting it to talk to your controller: https://www.arduino.cc/en/tutorial/blink

Then try this one to get the LED strip working: https://create.arduino.cc/projecthub/robocircuits/neopixel-tutorial-1ccfb9
 
Last edited:
For the tractor GPS receiver right now I'm going to use this receiver: http://racing.qstarz.com/Products/BT-Q818XT.html

Using it because I have it from past experiments. Its capable of 2m-ish accuracy at 5hz which isnt a lot but will do fine for testing. Also its battery powered and bluetooth so keeps the test install very simple from a cabling perspective.



The second prong of the gps system is the rtk base station. That will be based on a windows laptop connected to this receiver with an antenna: https://www.gnss.store/gnss-gps-mod...y.html#/27-add_antenna_ann_mb-without_antenna

You cant use any old receiver for the base, it has to be able to output raw RTCM data which not that many receivers can do. This seems to be one of the cheapest supported units to do it.

The base station laptop will initially just connect to a Wifi in the garden for testing. It will run Snip software: https://www.use-snip.com/ which acts as an NTRIP server (NTRIP is a network based RTK protocol)

To improve the functionality you would connect the laptop to the internet and have Snip relay to rtk2go.com which would publish your correction data which could then be accessed via a mobile hotspot connection on the tablet.
 
My plan is to get the lightbar working with the old receiver I have, then get the rtk working and see how good it makes things. After that I will add a roll sensor to compensate for slopes.

I've no major need for this. The few tonne of fert I throw out I manage not to overlap noticeably. I'm just doing it for the curiosity
 
To be honest you can read pretty much all I know about arduinos up this page.

I bought a nano because that's what the AgOpenGPS project uses
 
Back
Top